The expression of human recombinant proteins in heterologous cells has been well documented. Many production systems for recombinant proteins have become available, ranging from bacteria, yeasts, and fungi to insect cells, plant cells and mammalian cells. However, despite these developments, some production systems are still not optimal, or are only suited for production of specific classes of proteins. For instance, proteins that require post- or peri-translational modifications such as glycosylation, g-carboxylation, or g-hydroxylation cannot be produced in prokaryotic production systems. Another well-known problem with prokaryotic expression systems is the incorrect folding of the product to be produced, even leading to insoluble inclusion bodies in many cases.
Eukaryotic systems are an improvement in the production of, in particular, eukaryote derived proteins, but the available production systems still suffer from a number of drawbacks. The hypermannosylation in, for instance, yeast strains affects the ability of yeasts to properly express glycoproteins. Hypermannosylation often even leads to immune reactions when a therapeutic protein thus prepared is administered to a patient. Furthermore, yeast secretion signals are different from mammalian signals, leading to a more problematic transport of mammalian proteins, including human polypeptides, to the extracellular, which in turn results in problems with continuous production and/or isolation. Mammalian cells are widely used for the production of such proteins because of their ability to perform extensive post-translational modifications. The expression of recombinant proteins in mammalian cells has evolved dramatically over the past years, resulting in many cases in a routine technology.
In particular, Chinese hamster ovary cells (“CHO cells”) have become a routine and convenient production system for the generation of biopharmaceutical proteins and proteins for diagnostic purposes. A number of characteristics make CHO cells very suitable as a host cell. The production levels that can be reached in CHO cells are extremely high. The cell line provides a safe production system, which is free of infectious or virus-like particles. CHO cells have been extensively characterized, although the history of the original cell line is vague. CHO cells can grow in suspension until reaching high densities in bioreactors, using serum-free culture media; a dhfr-mutant of CHO cells (DG-44 clone, Urlaub et al., 1983) has been developed to obtain an easy selection system by introducing an exogenous dhfr gene and thereafter a well-controlled amplification of the dhfr gene and the transgene using methotrexate.
However, glycoproteins or proteins comprising at least two (different) subunits continue to pose problems. The biological activity of glycosylated proteins can be profoundly influenced by the exact nature of the oligosaccharide component. The type of glycosylation can also have significant effects on immunogenicity, targeting and pharmacokinetics of the glycoprotein. In recent years, major advances have been made in the cellular factors that determine the glycosylation, and many glycosyl transferase enzymes have been cloned. This has resulted in research aimed at metabolic engineering of the glycosylation machinery (Fussenegger et al., 1999; Lee et al., 1989; Vonach et al., 1998; Jenikins et al., 1998; Zhang et al., 1998; Muchmore et al., 1989). Examples of such strategies are described herein.
CHO cells lack a functional a-2,6 sialyl-transferase enzyme, resulting in the exclusive addition of sialyc acids to galactose via a-2,3 linkages. It is known that the absence of a-2,6 linkages can enhance the clearance of a protein from the bloodstream. To address this problem, CHO cells have been engineered to resemble the human glycan profile by transfecting the appropriate glycosyl transferases. CHO cells are also incapable of producing LewisX oligosaccharides. CHO cell lines have been developed that express human N-acetyl-D-glucosaminyltransferase and a-1,3-fucosyl-transferase III. In contrast, it is known that rodent cells, including CHO cells, produce CMP-N-acetylneuraminic acid hydrolase which glycosylates CMP-N-acetylneuraminic acids (Jenkins et al., 1996), an enzyme that is absent in humans. The proteins that carry this type of glycosylation can produce a strong immune response when injected (Kawashima et al., 1993). The recent identification of the rodent gene that encodes the hydrolase enzyme will most likely facilitate the development of CHO cells that lack this activity and will avoid this rodent-type modification.
Thus, it is possible to alter the glycosylation potential of mammalian host cells by expression of human glucosyl transferase enzymes. Yet, although the CHO-derived glycan structures on the recombinant proteins may mimic those present on their natural human counterparts, a potential problem exists in that they are still found to be far from identical. Another potential problem is that not all glycosylation enzymes have been cloned and are, therefore, available for metabolic engineering. The therapeutic administration of proteins that differ from their natural human counterparts may result in activation of the immune system of the patient and cause undesirable responses that may affect the efficacy of the treatment. Other problems using non-human cells may arise from incorrect folding of proteins that occurs during or after translation which might be dependent on the presence of the different available chaperone proteins. Aberrant folding may occur, leading to a decrease or absence of biological activity of the protein. Furthermore, the simultaneous expression of separate polypeptides that will together form proteins comprised of the different subunits, like monoclonal antibodies, in correct relative abundancies is of great importance. Human cells will be better capable of providing all necessary facilities for human proteins to be expressed and processed correctly.
It would thus be desirable to have methods for producing human recombinant proteins that involve a human cell that provides consistent human-type processing like post-translational and peri-translational modifications, such as glycosylation, which preferably is also suitable for large-scale production.